EN BUSCA DE ALTERNATIVAS DESARROLLADORAS

In the next two sections, I provide evidence for differentiating abstract from concrete rela- tional nouns. First, relational nouns in contexts of ‘property specification’ are shown to allow abstract but not concrete relational nouns. Second, the two types of relational noun are shown to differ with respect to how they are modified byactualand actually, making for a difference in potential CQ paraphrases. These issues are discussed in turn, but no attempt to incorporate these facts into the present approach is made.

3.6.2.1 CQs in Contexts of Specification

The construction in (143) contains the postverbal light quantifiersomething, which functions as a placeholder for a property or proposition (see Moltmann (2003) for discussion ofsome- thingin complement position). For lack of a better term, I call constructions which specify the content of the property or proposition immediately following the placeholder aproperty specificationenvironment. Curiously, a CQ from a Concrete Relational Noun is not admissi- ble in such environments (143a), even though an overt identity question (143b) is allowed. This pattern contrasts with CQs from Abstract Relational Nouns; both the CQ and the overt question are acceptable (143c–d).

(143) John knows something a. * the mayor of the town b. who the major of the town is

c. the price of milk

d. what the price of milk is

As a whole, the pattern observed in data in (144) contrasts with the pattern observed in (143). Here, only the individual interpretation of the DP is supported, and as such gender agreement prohibits the neuter noun phrase in (144c).14

(144) John knows someone: a. the mayor of the town

b. * who the major of the town is c. * the price of milk

d. * what the price of milk is

At first sight, the opposing patterns might be thought to be easily explained away by appeal to an ambiguous reading of something. It could be said that one reading supports the individual interpretation marked for neuter agreement, thereby explaining the contrast between (143a) and (143c). The other reading of something stands for proposition-like entities, and thus takes the full question of either Abstract or Concrete RNs. The account might continue that it is a curious fact of English that the animate someoneis not likewise ambiguous, lacking as it does the second, proposition-like interpretation.

Yet, this fact is still truly confounding; for we expect that DPs which function as CQs in other contexts should be interpreted in contexts that select for properties or propositions. Clearly, this is not possible with DPs denoting persons, or more accurately, with DPs denoting positions which select persons. Thus, any alleged ambiguity of something would not be sufficient to explain the contrast (143 – 144).

Similarly, the specificational pseudocleft exhibits a pattern identical to (143) (145) What John knows is

a. * the mayor of the town b. who the mayor of the town is

c. the price of milk

d. what the price of milk is

If an ambiguity for something is posited to explain the contrast in (143), the very same ambiguity must be operative in pseudoclefts (145). In particular, such a view would stipulate that the entire relative clause shares an ambiguity which was thought to be lexical in nature. Yet, these relative clauses are not lexical items, nor lexically fixed; any ambiguity proposed for pseudoclefts should be cashed out systematically, not on the basis of an exception.

3.6.2.2 The Interpretation ofactual

Although the paraphrase conditions of CQs are often treated as uniform across various types of noun phrases (cf. Jackendoff (1997)), certain abstract NPs diverge from concrete rela- tional NPs with respect to the interpretation of the intensional adjective actual.15 _{In brief,}

the adnominal modifieractualenforces different truth conditions on its complement than its adverbial counterpart actually, when they modify relational nouns. Furthermore, the para- phrasability conditions betweenabstractandconcreteare shown to differ. The interpretation of the question paraphrase withactualandactually reduce to the same interpretation if the noun is an abstract relational noun, but differ significantly if the noun is concrete. Thus, there is a choice of paraphrase between the adjective and the adverb in one type of relational noun, but not the other.

For illustration, suppose that John is an economist on the House Budget Committee and that the House Republicans have pushed a new bill that calls for a reduction of government spending for retirement. Being an expert on taxation and retirement, John recalculates the proposal under a more realistic economic model and finds that, lo and behold, his calcula- tions diverge from those quoted by the drafters of bill. Example (146) can be paraphrased

two ways, either as (146a) in which the adjectiveactualmodifiescost, or as (146b) in which case the adjective actual reinterpreted as an adverb, modifying the way in which the cost instantiates.

(146) John predicted the actual cost of the reform

a. John predicted what theactualcost of the reform was b. John predicted what the cost of the reformactuallywas

Do the truth conditions of (146a–b) differ? The difference, should it truly exist, is not entirely intuitive. Let S1 be a situation in which w′ is thought to be the actual world by

the addressee(s) of (146). In w′ the cost of the reform equals $26 million dollars. John’s recalculations show that the cost of the reform inw′′is instead $28 million dollars.

In this situation, example (146) has the force of asserting that w′′, and not w′, is the actual world with respect to the cost of the reform. It seems that either (146a–b) captures this reading.

Let S2 be the situation in which John has uncovered a case of attempted graft in which

the monies from the proposed reform was, in reality, earmarked for kickbacks to government 15_{Actualhas many uses, including one for signalling negation or revision in discourse. It is important not}

to interpret the examples with such a use. Other intensional-like adjectives such asreal, suffer from the same difficulties in interpretation. Another candidate is the non-intersective alleged. However, it might invoke a hidden speech act, say from an implicit accuser, which may degrade the intended adverbial reading. At any rate, the productivity of such examples is not at all certain. See also Zimmerman (2000) for an account of

occasionalthat displays a similar type of ambiguity, although on its adverbial reading occasionallyno longer modifies the DP, but instead the entire sentence. Further investigation is required to determine the relationship between these cases.

World Reform cost

w′ 7→ $26m

w′′ 7→ $28m

Table 3.5: Cost Concept in situation S1

contractors. John may not even be able to calculate the exact number, and yet might still

predict that the cost of the reform is always equal to the funds diverted towards another fund (or even a part thereof). John need not know or predict any particular number, only a relation that sufficiently establishes (a relative) identity. In this situation, the cost in dollars of the reform might not differ between possible worlds, sayw′andw′′, but the way in which John identifies it.

Worlds Reform cost

w′ 7→ $26m

w′′ 7→ $26m

Worlds Kickback amount

w′ 7→ $10m

w′′ 7→ $26m

Table 3.6: Cost concept in situation S2

The point is that in S2 example (146) seems to require not that the costs inw′ andw′′ differ

and thatw′′is the actual world, but rather that the cost of the reform shares a property or is identical with another concept in the actual world. In this situation, w′ would be ruled out on these grounds.

It is important to determine whether both paraphrases (146a–b) are appropriate in situ- ation S2. While it is not completely clear to me, I am strongly inclined to suggest that they

are. That is, I maintain that there is no crucial difference between the question paraphrases in the CQ context.16 _{So far, this conclusion is unsurprising if we accept the common view}

thatactualandactuallyare mere syntactic variants of one another. The contrast that appears with concrete relational nouns clearly makes this view untenable.

Suppose that John has been investigating a vast conspiracy instituting a shadow gov- ernment in America. He uncovers that the presidency is, in reality, occupied by another politician. In this scenario, (147) can only be paraphrased as (147a).

(147) John discovered the actual president

a. John discovered who/what the actual president was b. # John discovered who/what the president actually was

16_{The slight difference could be cashed out as follows: it seems that (146a) might be appropriately used}

when there was a slight miscalculation and theactual costis another value. In contrast, (146b) has a stronger reading, namely that the quality constituting the cost is different in the actual world then expected. In such an instance, the numerical value could be the same or not, but the character of the way in which it constitutes a cost must differ.

To see the difference, suppose that S3 is the situation in which the addressee(s) of (147)

are thought to believe that they are in w′ and the president (of the United Sates) in the actual world denotes some individual: George W. Bush. However, John discovers the shadow government’s plot, in which the president in w′ is not president at all, but rather a political puppet. Some other individual – say Dick Cheney – is in fact president, as John discovers.

World President

w′ _{7→ GWB}

w′′ 7→ DC

Table 3.7: President concept in situation S3

It is a striking contrast to (146) that the paraphrase in (147b), containing the adverb

actually, is not an appropriate paraphrase. By itself, (147b) means something quite different than either (147) or (147a). Let S₄ be a situation in which John discovers that in reality the president occupies another position, such as a renegade semanticist, say Aaron Aaronson, who writes mysterious yet brilliant papers anonymously. In this very counterfactual world, it is not the referent of the president that changes. Instead, the extension of the president concept is equivalent to the extension of the renegade semanticist concept in whatever world is actual.

Worlds President

w′ 7→ GWB

w′′ 7→ GWB

Worlds Renegade Semanticist

w′ 7→ AA

w′′ 7→ GWB

Table 3.8: Cost concept in situation S4

In this case, we could utter (147b), but not (147) or (147a). The point is that while (147b) is felicitous on other occasions of use, it is not a readily available paraphrase of (147). The difference in interpretation here is surprisingly robust, especially in contrast to the pattern illustrated by (146).

The contrast between (146) and (147) suggests that CQs containing actual and an ab- stract relational noun may be paraphrased in either (a) the adverbial reading or (b) the adjectival reading ofactual, where as those CQs containing concrete relational nouns do not license the adverbial paraphrase under similar circumstances.

Note that this contrast is not limited to CQ environments. The basic pattern is replicated in simple matrix clauses, such as those below.

(148) ABSTRACT RELATIONAL NOUN

a. The actual cost of the reform is $26m b. The cost of the reform is actually $26m

(149) CONCRETE RELATIONAL NOUN

a. The actual president is Dick Cheney b. The president is actually Dick Cheney

Although the different senses of (148a–b) is subtle, the distinct interpretations of (149a–b) is non-trivial. While (149a) claims that the person who truly holds the presidency is not Bush, but Cheney, (149b) claims something much stronger. It means that the referent of the term

the president, George Bush, is in fact Dick Cheney. So, it appears that the relation of identity which is asserted by (149a) is about who holds office in the actual world – a world implicitly other than how we had imagined. Example (149b), on the other hand, is about the identity of the person who holds office in the worlds we believe to be actual. This is perhaps why (149b) sounds like a strange assertion; it equates two objects we believe to be distinct.

The pattern is again replicated with descriptions in the post-copular position: (150) ABSTRACT RELATIONAL NOUN

a. The actual cost of the reform is the amount of the kickback b. The cost of the reform is actually the amount of the kickback (151) CONCRETE RELATIONAL NOUN

a. The actual president is the man in the brown hat b. The president is actually the man in the brown hat

The full ramifications of this contrast are far from clear. It is possible that the difference between types of relational noun could be fomulated within a two-dimensional semantic theory (see Chalmers (2007) for an excellent overview article, as well as Soames (2005)), in which the adverbialactuallymodifies the 1-intension, or the standard Carnap/Fregean sense of an expression, whereas the adjective actual modifies the 2-intension, which determines how the reference of the term is fixed. All the same, the exact solution of this puzzle is left as a challenge for future, and more comprehensive, theories of concealed question. Before ending this chapter, we briefly review the material discussed above.